Update - We are experiencing intermittent network outages.
We try to resolve these as soon as possible, but cannot anticipate and prevent the outages at this time.
Outages last about 5 minutes at a frequency of roughly once per hour.
We are preparing for our scheduled maintenance to address the network outages, which will take place at 00:30 - 01:30 this evening
Feb 18, 2026 - 21:58 SAST
We try to resolve these as soon as possible, but cannot anticipate and prevent the outages at this time.
Outages last about 5 minutes at a frequency of roughly once per hour.
We are preparing for our scheduled maintenance to address the network outages, which will take place at 00:30 - 01:30 this evening
Feb 18, 2026 - 21:58 SAST
Update - We are continuing to work on a fix for this issue.
Feb 18, 2026 - 21:07 SAST
Feb 18, 2026 - 21:07 SAST
Update - We are continuing to work on a fix for this issue.
Feb 18, 2026 - 20:57 SAST
Feb 18, 2026 - 20:57 SAST
Update - Network interruptions are ongoing.
We continue to work at repairing and minimizing the impact.
Outside connectivity to VMs has dropped at 20:35 and we anticipate the network to recover in the next few minutes
Feb 18, 2026 - 20:38 SAST
We continue to work at repairing and minimizing the impact.
Outside connectivity to VMs has dropped at 20:35 and we anticipate the network to recover in the next few minutes
Feb 18, 2026 - 20:38 SAST
Identified - Following the network outage 19:35-19:41 we need to manually start our cluster services.
Functionality on app.cloudafrica.net will be severely limited.
We have scheduled maintenance later tonight to deal with this and thank you for your understanding and apologise for the interruptions.
Feb 18, 2026 - 19:43 SAST
Functionality on app.cloudafrica.net will be severely limited.
We have scheduled maintenance later tonight to deal with this and thank you for your understanding and apologise for the interruptions.
Feb 18, 2026 - 19:43 SAST
Investigating - Starting at 19:35 we received alerts that our network is unreachable.
User will have lost connectivity to their VMs.
We are working to resolve connectivity and will update as soon as progress is made
Feb 18, 2026 - 19:37 SAST
User will have lost connectivity to their VMs.
We are working to resolve connectivity and will update as soon as progress is made
Feb 18, 2026 - 19:37 SAST
Update - Cluster services are up and the network is stable at the moment.
Functionality on app.cloudafrica.net has been restored and all traffic is flowing normally.
We continue to monitor until the maintenance scheduled for tonight.
Feb 18, 2026 - 15:18 SAST
Functionality on app.cloudafrica.net has been restored and all traffic is flowing normally.
We continue to monitor until the maintenance scheduled for tonight.
Feb 18, 2026 - 15:18 SAST
Update - Following a network outage we experienced a slow and degraded network flow, users may have seen packet loss on their connections.
The network is operational with no packet losses, and all VMs are operational and accessible over the internet.
Cluster services are in the process of restarting, and functionality on app.cloudafrica.net is degraded until the restart is complete.
Feb 18, 2026 - 13:15 SAST
The network is operational with no packet losses, and all VMs are operational and accessible over the internet.
Cluster services are in the process of restarting, and functionality on app.cloudafrica.net is degraded until the restart is complete.
Feb 18, 2026 - 13:15 SAST
Update - The app.cloudafrica.net platform is reachable and all VMs are responding to network traffic.
Some users may experience packet loss at the moment.
Our cluster services are also down and app.cloudafrica.net functionality is severely limited.
Feb 18, 2026 - 12:36 SAST
Some users may experience packet loss at the moment.
Our cluster services are also down and app.cloudafrica.net functionality is severely limited.
Feb 18, 2026 - 12:36 SAST
Update - The network has gone offline and is unreachable from outside of Cloudafrica.
All services are disrupted and unreachable.
We are actively working on resolving this.
Feb 18, 2026 - 12:32 SAST
All services are disrupted and unreachable.
We are actively working on resolving this.
Feb 18, 2026 - 12:32 SAST
Update - All services have been up for a short while, however, we are noticing packet losses on our network which effects VM performance and my look like there is a problem with a VM, we are investigating this and treating it as urgently as other incidents.
Feb 18, 2026 - 12:10 SAST
Feb 18, 2026 - 12:10 SAST
Monitoring - Cluster services have been restarted and are running at the moment.
All functionality on our platform has been restored.
We will continue to monitor and investigate in an attempt to prevent future incidents.
Feb 18, 2026 - 10:26 SAST
All functionality on our platform has been restored.
We will continue to monitor and investigate in an attempt to prevent future incidents.
Feb 18, 2026 - 10:26 SAST
Update - We are busy restarting cluster services now.
Functionality on app.cloudafrica.net will be unavailable during this time, however, all VMs are fully operational and reachable.
functionality that is unavailable at the moment include: stop, start, reboot shutdown, backup, firewall rule management, rename & restore
Feb 18, 2026 - 09:53 SAST
Functionality on app.cloudafrica.net will be unavailable during this time, however, all VMs are fully operational and reachable.
functionality that is unavailable at the moment include: stop, start, reboot shutdown, backup, firewall rule management, rename & restore
Feb 18, 2026 - 09:53 SAST
Update - After the network outage that between 08:34 - 08:56, our cluster services failed to recover.
We are restarting these services now.
Functionality on app.cloudafrica.net will be unavailable during this time, however, all VMs are fully operational and reachable.
functionality that is unavailable at the moment include: stop, start, reboot shutdown, backup, firewall rule management, rename & restore
Feb 18, 2026 - 09:08 SAST
We are restarting these services now.
Functionality on app.cloudafrica.net will be unavailable during this time, however, all VMs are fully operational and reachable.
functionality that is unavailable at the moment include: stop, start, reboot shutdown, backup, firewall rule management, rename & restore
Feb 18, 2026 - 09:08 SAST
Identified - The network has been restored and all customer VMs are accessible via ssh or rdp connections.
Currently functionality on app.cloudafrica.net is not available, stop, start, rename, firewall management and backups are unavailable at the moment.
Feb 18, 2026 - 08:59 SAST
Currently functionality on app.cloudafrica.net is not available, stop, start, rename, firewall management and backups are unavailable at the moment.
Feb 18, 2026 - 08:59 SAST
Investigating - We have a major network outage affecting all traffic to our network.
Customer VMs and our various platforms are currently inaccessible
Feb 18, 2026 - 08:52 SAST
Customer VMs and our various platforms are currently inaccessible
Feb 18, 2026 - 08:52 SAST
Web Sites
Degraded Performance
API
Degraded Performance
Storage Services
Degraded Performance
Cloud Services
Degraded Performance
Network
Partial Outage
Operational
Degraded Performance
Partial Outage
Major Outage
Maintenance
Major outage
Partial outage
No downtime recorded on this day.
No data exists for this day.
had a major outage.
had a partial outage.
Scheduled Maintenance
Subject: Emergency Network Maintenance - Thursday 19th February 2026 at 00H30-01H30 SAST / UTC+2 Feb 19, 2026 00:30-01:30 SAST
We're performing emergency network maintenance tonight to address what we believe to be the root cause of recent service interruptions.
MAINTENANCE DETAILS:
Date: February 19, 2026
Time: 00H30 to 01H30 SAST (UTC+2)
Duration: 60-minute window (periods of 30-90 seconds downtime during this window)
Impact: Brief service interruptions, automatic reconnection
Scope: All services in Teraco JB1 East, Teraco JB1 West, and OADC JNB1
THE PROBLEM:
Over the past 16 days, we've experienced 15 unplanned service interruptions totaling 72 minutes of downtime, with 4 incidents occurring today alone (32 minutes of collective downtime). These outages have been characterized by CPU saturation on core switches during traffic peaks, leading to packet drops and service degradation.
TROUBLESHOOTING:
Initially, we believed these issues might be firmware-related. We completed comprehensive firmware upgrades last night across all 15 core switches in our network fabric, which interconnects at 200Gbps across Teraco JHB1 East, Teraco JHB1 West, and OADC JNB1 datacenters. While these upgrades improved stability marginally, the core outage pattern persisted.
After significant further analysis of switch logs, CPU utilization patterns, and packet drop statistics, it appears that the root cause is excessive protocol overhead in our MSTP (Multiple Spanning Tree Protocol) configuration. Our switches are running MSTP with ~4000 VLANs, forcing the CPU to perform complex MD5 hash calculations and region verification every 2 seconds. This consumes 10-15% baseline CPU and spikes to 100% during traffic bursts, at which point packet forwarding degrades and outages occur.
SOLUTION:
Based on this analysis, we're migrating to RSTP (Rapid Spanning Tree Protocol), which:
- Provides identical network protection and redundancy - and considering that we do not use the features that MSTP is designed for
- Eliminates unnecessary protocol overhead (3-5x reduction)
- Frees CPU resources to handle traffic instead of redundant calculations
- Aligns with industry best practices for our network topology
This change is analogous to disabling resource-intensive features that provide no functional benefit - we maintain the same protection with significantly lower overhead.
EXPECTED RESULTS:
If our analysis is correct (which we believe it is), this change should deliver:
- 50-70% reduction in CPU utilization during traffic peaks
- Significantly improved stability and fewer service interruptions
- Faster packet processing and better overall performance
- Elimination of the outage pattern we've observed
ONGOING ANALYSIS:
Between now and the maintenance window (00:30 tonight), we will continue conducting detailed core switch reviews to verify this diagnosis and ensure we're addressing the root cause. We're monitoring CPU patterns, packet drops, queue depths, and protocol overhead in real-time.
While we're confident in our analysis, we're scheduling this emergency maintenance on the basis that MSTP overhead is the primary issue, while maintaining vigilance for any additional contributing factors.
TECHNICAL ASSURANCE:
- No change to network topology, routing, or VLAN configuration is planned
- Identical loop prevention and failover capabilities
- Industry-standard protocol (IEEE 802.1w) used globally in production
- Comprehensive rollback plan ready if needed (though not expected)
- Change will be applied systematically across affected switches
WHY EMERGENCY TIMING:
With the frequency and severity of outages increasing (4 incidents today), the current configuration is clearly actively degrading service quality. Our analysis indicates each day of delay adds approximately 4-5 minutes of unplanned downtime.
POST-MAINTENANCE:
We will monitor network performance continuously for 48 hours following the change. If the outage pattern continues despite this optimization, we'll immediately escalate to additional diagnostic measures and corrective actions.
You'll receive a completion notice with initial results shortly after the maintenance window closes.
We sincerely apologize for the short notice and any inconvenience. Your service reliability is our highest priority, and we're taking systematic, evidence-based action to restore the stability you expect from us.
Posted on Feb 18, 2026 - 15:16 SAST
MAINTENANCE DETAILS:
Date: February 19, 2026
Time: 00H30 to 01H30 SAST (UTC+2)
Duration: 60-minute window (periods of 30-90 seconds downtime during this window)
Impact: Brief service interruptions, automatic reconnection
Scope: All services in Teraco JB1 East, Teraco JB1 West, and OADC JNB1
THE PROBLEM:
Over the past 16 days, we've experienced 15 unplanned service interruptions totaling 72 minutes of downtime, with 4 incidents occurring today alone (32 minutes of collective downtime). These outages have been characterized by CPU saturation on core switches during traffic peaks, leading to packet drops and service degradation.
TROUBLESHOOTING:
Initially, we believed these issues might be firmware-related. We completed comprehensive firmware upgrades last night across all 15 core switches in our network fabric, which interconnects at 200Gbps across Teraco JHB1 East, Teraco JHB1 West, and OADC JNB1 datacenters. While these upgrades improved stability marginally, the core outage pattern persisted.
After significant further analysis of switch logs, CPU utilization patterns, and packet drop statistics, it appears that the root cause is excessive protocol overhead in our MSTP (Multiple Spanning Tree Protocol) configuration. Our switches are running MSTP with ~4000 VLANs, forcing the CPU to perform complex MD5 hash calculations and region verification every 2 seconds. This consumes 10-15% baseline CPU and spikes to 100% during traffic bursts, at which point packet forwarding degrades and outages occur.
SOLUTION:
Based on this analysis, we're migrating to RSTP (Rapid Spanning Tree Protocol), which:
- Provides identical network protection and redundancy - and considering that we do not use the features that MSTP is designed for
- Eliminates unnecessary protocol overhead (3-5x reduction)
- Frees CPU resources to handle traffic instead of redundant calculations
- Aligns with industry best practices for our network topology
This change is analogous to disabling resource-intensive features that provide no functional benefit - we maintain the same protection with significantly lower overhead.
EXPECTED RESULTS:
If our analysis is correct (which we believe it is), this change should deliver:
- 50-70% reduction in CPU utilization during traffic peaks
- Significantly improved stability and fewer service interruptions
- Faster packet processing and better overall performance
- Elimination of the outage pattern we've observed
ONGOING ANALYSIS:
Between now and the maintenance window (00:30 tonight), we will continue conducting detailed core switch reviews to verify this diagnosis and ensure we're addressing the root cause. We're monitoring CPU patterns, packet drops, queue depths, and protocol overhead in real-time.
While we're confident in our analysis, we're scheduling this emergency maintenance on the basis that MSTP overhead is the primary issue, while maintaining vigilance for any additional contributing factors.
TECHNICAL ASSURANCE:
- No change to network topology, routing, or VLAN configuration is planned
- Identical loop prevention and failover capabilities
- Industry-standard protocol (IEEE 802.1w) used globally in production
- Comprehensive rollback plan ready if needed (though not expected)
- Change will be applied systematically across affected switches
WHY EMERGENCY TIMING:
With the frequency and severity of outages increasing (4 incidents today), the current configuration is clearly actively degrading service quality. Our analysis indicates each day of delay adds approximately 4-5 minutes of unplanned downtime.
POST-MAINTENANCE:
We will monitor network performance continuously for 48 hours following the change. If the outage pattern continues despite this optimization, we'll immediately escalate to additional diagnostic measures and corrective actions.
You'll receive a completion notice with initial results shortly after the maintenance window closes.
We sincerely apologize for the short notice and any inconvenience. Your service reliability is our highest priority, and we're taking systematic, evidence-based action to restore the stability you expect from us.
Posted on Feb 18, 2026 - 15:16 SAST